[MLIR][GPU] Add error checking to loop.parallel to gpu transform.

[herhut]

Summary

Instead of crashing on malformed input, the pass now produces error
messages.

[ftynse]

A couple of nits, looks good otherwise.

mlir/lib/Conversion/LoopsToGPU/LoopsToGPU.cpp

575

Nit: DenseMap is re-exported to the mlir namespace

635

isa_and_nonnull here and below

652

Nit: I think you can do return paralellOp.emitOpError() ..., it is convertible to LogicalResult

675

Typo: processor

[herhut]

Review comments.

[herhut]

Thanks!

[rriddle]

mlir/lib/Conversion/LoopsToGPU/LoopsToGPU.cpp

761

Drop the llvm:: here please.

810

nit: Can we use a type instead of auto? It helps with readability.

mlir/lib/Conversion/LoopsToGPU/LoopsToGPU.cpp

 /// To note the end of the current scope in case a loop or conditional was
 /// inserted, a sentinel (the `gpu.launch` operation) is inserted into the
 /// worklist. This signals the processor of the worklist to pop the rewriter
 /// one scope-level up.
 static LogicalResult processParallelLoop(ParallelOp parallelOp,
                                          gpu::LaunchOp launchOp,
                                          BlockAndValueMapping &cloningMap,
                                          SmallVectorImpl<Operation *> &worklist,
+                                         llvm::DenseMap<int, Value> &bounds,
                                          PatternRewriter &rewriter) {
   // TODO(herhut): Verify that this is a valid GPU mapping.
   // processor ids: 0-2 block [x/y/z], 3-5 -> thread [x/y/z], 6-> sequential
   ArrayAttr mapping =
       parallelOp.getAttrOfType<ArrayAttr>(gpu::kMappingAttributeName);
 
   // TODO(herhut): Support reductions.
   if (!mapping || parallelOp.getNumResults() != 0)
if (annotation.processor < gpu::LaunchOp::kNumConfigOperands) {
         // We pass as the single opererand to the bound-map the number of
         // iterations, which is (upperBound - lowerBound) ceilDiv step. To
         // support inner loops with dynamic upper bounds (as generated by e.g.
         // tiling), try to derive a max for the bounds. If the used bound for
         // the hardware id is imprecise, wrap the contained code into a
         // conditional. If the lower-bound is constant or defined before the
         // launch, we can use it in the launch bounds. Otherwise fail.
         if (!launchIndependent(lowerBound) &&
-            !isa<ConstantOp>(lowerBound.getDefiningOp()))
+            !dyn_cast_or_null<ConstantOp>(lowerBound.getDefiningOp()))
           return failure();
         // The step must also be constant or defined outside of the loop nest.
-        if (!launchIndependent(step) && !isa<ConstantOp>(step.getDefiningOp()))
+        if (!launchIndependent(step) &&
+            !dyn_cast_or_null<ConstantOp>(step.getDefiningOp()))
           return failure();
         // If the upper-bound is constant or defined before the launch, we can
         // use it in the launch bounds directly. Otherwise try derive a bound.
-        bool boundIsPrecise = launchIndependent(upperBound) ||
-                              isa<ConstantOp>(upperBound.getDefiningOp());
+        bool boundIsPrecise =
+            launchIndependent(upperBound) ||
+            dyn_cast_or_null<ConstantOp>(upperBound.getDefiningOp());
         {
           PatternRewriter::InsertionGuard guard(rewriter);
           rewriter.setInsertionPoint(launchOp);
           if (!boundIsPrecise) {
             upperBound = deriveStaticUpperBound(upperBound, rewriter);
-            if (!upperBound)
+            if (!upperBound) {
+              parallelOp.emitOpError() << "cannot derive loop-invariant upper "
+                                          "bound for number of iterations";
               return failure();
             }
+          }
           // Compute the number of iterations needed. We compute this as an
           // affine expression ceilDiv (upperBound - lowerBound) step. We use
           // affine.apply here so that it composes nicely with the provided map.
           AffineMap stepMap =
               AffineMap::get(0, 3,
                              ((rewriter.getAffineSymbolExpr(0) -
-                              rewriter.getAffineSymbolExpr(1)).ceilDiv(
-                                  rewriter.getAffineSymbolExpr(2))));
+                               rewriter.getAffineSymbolExpr(1))
+                                  .ceilDiv(rewriter.getAffineSymbolExpr(2))));
           Value launchBound = rewriter.create<AffineApplyOp>(
               loc, annotation.boundMap.compose(stepMap),
               ValueRange{
                   ensureLaunchIndependent(
                       cloningMap.lookupOrDefault(upperBound)),
                   ensureLaunchIndependent(
                       cloningMap.lookupOrDefault(lowerBound)),
                   ensureLaunchIndependent(cloningMap.lookupOrDefault(step))});
-          launchOp.setOperand(annotation.processor, launchBound);
+          if (bounds.find(annotation.processor) != bounds.end()) {
+            parallelOp.emitOpError()
+                << "cannot redefine the bound for precessor "
+                << annotation.processor;
+            return failure();
+          }
+          bounds[annotation.processor] = launchBound;
         }
         if (!boundIsPrecise) {
           // We are using an approximation, create a surrounding conditional.
           Value originalBound = std::get<3>(config);
           CmpIOp pred = rewriter.create<CmpIOp>(
               loc, CmpIPredicate::slt, newIndex,
               cloningMap.lookupOrDefault(originalBound));
           loop::IfOp ifOp = rewriter.create<loop::IfOp>(loc, pred, false);
ParallelToGpuLaunchLowering::matchAndRewrite(ParallelOp parallelOp,
   gpu::LaunchOp launchOp = rewriter.create<gpu::LaunchOp>(
       parallelOp.getLoc(), constantOne, constantOne, constantOne, constantOne,
       constantOne, constantOne);
   rewriter.setInsertionPointToEnd(&launchOp.body().front());
   rewriter.create<gpu::TerminatorOp>(loc);
   rewriter.setInsertionPointToStart(&launchOp.body().front());
 
   BlockAndValueMapping cloningMap;
+  llvm::DenseMap<int, Value> launchBounds;
   SmallVector<Operation *, 16> worklist;
   if (failed(processParallelLoop(parallelOp, launchOp, cloningMap, worklist,
-                                 rewriter)))
+                                 launchBounds, rewriter)))
     return matchFailure();
 
   // Whether we have seen any side-effects. Reset when leaving an inner scope.
   bool seenSideeffects = false;
   // Whether we have left a nesting scope (and hence are no longer innermost).
   bool leftNestingScope = false;
   while (!worklist.empty()) {
     Operation *op = worklist.pop_back_val();
     // Now walk over the body and clone it.
     // TODO: This is only correct if there either is no further loop.parallel
     //       nested or this code is side-effect free. Otherwise we might need
     //       predication. We are overly consertaive for now and only allow
     //       side-effects in the innermost scope.
     if (auto nestedParallel = dyn_cast<ParallelOp>(op)) {
       // Before entering a nested scope, make sure there have been no
       // sideeffects until now.
       if (seenSideeffects)
         return matchFailure();
       // A nested loop.parallel needs insertion of code to compute indices.
       // Insert that now. This will also update the worklist with the loops
       // body.
-      processParallelLoop(nestedParallel, launchOp, cloningMap, worklist,
-                          rewriter);
+      if (failed(processParallelLoop(nestedParallel, launchOp, cloningMap,
+                                     worklist, launchBounds, rewriter)))
+        return matchFailure();
     } else if (op == launchOp.getOperation()) {
       // Found our sentinel value. We have finished the operations from one
       // nesting level, pop one level back up.
       auto parent = rewriter.getInsertionPoint()->getParentOp();
       rewriter.setInsertionPointAfter(parent);
       leftNestingScope = true;
       seenSideeffects = false;
     } else {
       // Otherwise we copy it over.
       Operation *clone = rewriter.clone(*op, cloningMap);
       cloningMap.map(op->getResults(), clone->getResults());
       // Check for side effects.
       seenSideeffects |= !clone->hasNoSideEffect();
       // If we are no longer in the innermost scope, sideeffects are disallowed.
       if (seenSideeffects && leftNestingScope)
         return matchFailure();
     }
   }
 
+  // Now that we succeeded creating the launch operation, also update the
+  // bounds.
+  for (auto bound : launchBounds)
+    launchOp.setOperand(std::get<0>(bound), std::get<1>(bound));
+
   rewriter.eraseOp(parallelOp);
   return matchSuccess();
 }
 
 void mlir::populateParallelLoopToGPUPatterns(OwningRewritePatternList &patterns,
                                              MLIRContext *ctx) {
   patterns.insert<ParallelToGpuLaunchLowering>(ctx);
 }

mlir/test/Conversion/LoopsToGPU/parallel_loop.mlir

-// RUN: mlir-opt -convert-parallel-loops-to-gpu -split-input-file %s | FileCheck %s -dump-input-on-failure
+// RUN: mlir-opt -convert-parallel-loops-to-gpu -split-input-file -verify-diagnostics %s | FileCheck %s -dump-input-on-failure
 
 // 2-d parallel loop mapped to block.y and block.x
 
 func @parallel_loop_bidy_bidx(%arg0 : index, %arg1 : index, %arg2 : index,
                               %arg3 : index, %arg4 : index, 
                               %buf : memref<?x?xf32>,
                               %res : memref<?x?xf32>) {
   %step = constant 2 : index

 // CHECK:                 store [[VAL_52]], [[VAL_44]]{{\[}}[[VAL_45]], [[VAL_47]]] : memref<?x?xf32, #[[MAP5]]>
 // CHECK:               }
 // CHECK:             }
 // CHECK:             gpu.terminator
 // CHECK:           }
 // CHECK:           return
 // CHECK:         }
 // CHECK:       }
+
+// -----
+
+// Mapping to the same processor twice.
+
+func @parallel_double_map(%arg0 : index, %arg1 : index, %arg2 : index,
+                          %arg3 : index,
+                          %buf : memref<?x?xf32>,
+                          %res : memref<?x?xf32>) {
+  %four = constant 4 : index
+  // expected-error@+2 {{cannot redefine the bound for precessor 1}}
+  // expected-error@+1 {{failed to legalize operation 'loop.parallel'}}
+  loop.parallel (%i0, %i1) = (%arg0, %arg1) to (%arg2, %arg3)
+                                          step (%four, %four)  {
+  } { mapping = [
+      {processor = 1, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>},
+      {processor = 1, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>}
+    ] }
+  return
+}
+
+// -----
+
+// Loop with loop-variant upper bound.
+
+func @parallel_loop_loop_variant_bound(%arg0 : index, %arg1 : index, %arg2 : index,
+                                       %arg3 : index,
+                                       %buf : memref<?x?xf32>,
+                                       %res : memref<?x?xf32>) {
+  %zero = constant 0 : index
+  %one = constant 1 : index
+  %four = constant 4 : index
+  // expected-error@+1 {{failed to legalize operation 'loop.parallel'}}
+  loop.parallel (%i0, %i1) = (%arg0, %arg1) to (%arg2, %arg3)
+                                          step (%four, %four)  {
+    // expected-error@+1 {{cannot derive loop-invariant upper bound}}                                        
+    loop.parallel (%si0, %si1) = (%zero, %zero) to (%i0, %i1)
+                                            step (%one, %one)  {
+      %idx0 = addi %i0, %si0 : index
+      %idx1 = addi %i1, %si1 : index
+      %val = load %buf[%idx0, %idx1] : memref<?x?xf32>
+      store %val, %res[%idx1, %idx0] : memref<?x?xf32>
+    } { mapping = [
+        {processor = 4, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>},
+        {processor = 6, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>}
+      ] }
+  } { mapping = [
+      {processor = 1, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>},
+      {processor = 6, map = affine_map<(d0) -> (d0)>, bound = affine_map<(d0) -> (d0)>}
+    ] }
+  return
+}